Emotion detection in psychological texts by fine-tuning BERT using emotion–cause pair extraction

International Journal of Speech Technology - The task of detecting and recognizing the state of emotion when dealing with textual data in the machine learning domain is not a straightforward path....     

Spatiotemporal dynamics induced by delay and diffusion in a predator–prey model with mutual interference among the predator

In this work, we study a reaction–diffusion predator–prey model with mutual interference among the predators while searching for food. We prove that the model exhibits bistability, which indicates that there are no patterns for our model. When time delay is incorporated into the model, multiple stability switches phenomenon of positive constant steady state emerged. By taking delay as a bifurcation parameter, the Hopf bifurcations at the positive constant steady state are proved to occur for a sequence of critical values of the delay. The algorithm for determining the direction and the stability of the bifurcating periodic solutions is also derived. The delay–diffusion driven Turing instability of the positive constant steady state is investigated. Our results show that delay and diffusion can create periodic oscillatory patterns of spatially homogeneous and inhomogeneous and Turing patterns.     

A capital budgeting problem for preventing workplace mobbing by using analytic hierarchy process and fuzzy 0–1 bidimensional knapsack model

In this paper, a capital budgeting problem for preventive measures of workplace mobbing based on fuzzy 0–1 bidimensional knapsack model with non-financial and financial budget limits is proposed. The weights to be used as the objective function coefficients of the model are obtained from analytic hierarchy process (AHP) methodology that incorporates possible causes of workplace mobbing on criteria level, and possible preventive measures on alternatives level of AHP hierarchy. The quantification of non-financial and financial budgets as well as non-financial and financial costs is developed based on their relative weights. In deterministic model, the relative weights are directly used, whereas in fuzzy model, they are quantified. The defuzzification of the fuzzy model is proposed to be made by using t-norm and t-conorm fuzzy relations which are expected to give the most optimistic, and the most pessimistic results, respectively. The results of the hypothetical example verify the expectations.     

How consumer innovativeness,technological expertise,and consideration set size can explain mobile commerce use: An extended understanding using a moderation–mediation model

Information Systems and e-Business Management - This study integrates consumer innovativeness (CI) and technological expertise (TE) in consumer attitudes and mobile commerce use (MCU) and...     

Reliable estimation of evapotranspiration on agricultural fields predicted by the Priestley–Taylor model using soil moisture data from ground and remote sensing observations compared with the Common Land Model

Evapotranspiration (ET) is a crucial factor in understanding the hydrological cycle and is essential to many applications in hydrology, ecology and water resources management. However, reliable ET measurements and predictions for a range of temporal and spatial scales are difficult. This study focused on the comparison of ET estimates using a relatively simple model, the Priestley–Taylor (P-T) approach, and the physically based Common Land Model (CLM) using ground and remotely sensed soil moisture data as input. The results from both models were compared directly with hourly eddy covariance measurements at two agricultural field sites during the Soil Moisture–Atmosphere Coupling Experiment (SMACEX) in the corn soybean production region in the Upper Midwest, USA. The P-T model showed a significant overestimation of the potential ET compared to the measurements, with a root mean square error (RMSE) between 115 and 130 W m^–2. Actual ET was better predicted by the CLM, with the RMSE ranging between 50 and 75 W m^–2. However, actual ET from the P-T model constrained with a soil moisture dependency parameterization showed improved results when compared to the measurements, with a significantly reduced bias and RMSE values between 60 and 65 W m^–2. This study suggests that even with a simple semi-empirical ET model, similar performance in estimating actual ET for agricultural crops compared to more complex land surface–atmosphere models (i.e. the CLM) can be achieved when constrained with the soil moisture function. This suggests that remote sensing soil moisture estimates from the Advanced Microwave Scanning Radiometer – Earth Observing System (AMSR-E) and others such as the Soil Moisture and Ocean Salinity (SMOS) mission may be effective alternatives under certain environmental conditions for estimating actual ET of agricultural crops using a fairly simple algorithm.     

A regional algorithm for separating light absorption by chlorophyll-a and coloured detrital matter in the Black Sea,using 480–560 nm bands from ocean colour scanners

The aim of this study is to modify the regional algorithm for Moderate Resolution Imaging Spectroradiometer (MODIS) and Medium-spectral Resolution Imaging Spectrometer (MERIS) bands using newly available data of seasonal and spatial variability of light absorption by all optically active components in the Black Sea, and to obtain a merged product based on data retrieved from all the colour scanners that have operated since September 1997. Comparison of chlorophyll-a concentration (chl-a) simulated by the standard National Aeronautics and Space Administration (NASA) algorithm with in situ chl-a measurements showed that the NASA algorithm provided incorrect chl-a assessment of Black Sea shelf and deep-sea waters during spring?summer. Originally the standard NASA algorithm could be applied if there was a high correlation between light absorption by phytoplankton (a_ph) and that by coloured dissolved and suspended organic matter (a_CDM), which is not the case in the Black Sea. Consequently, development of the correct regional chl-a algorithm requires splitting of light absorption into a_ph and a_CDM. This issue has been resolved by the proposed regional algorithm developed for the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) using remote-sensing reflectance in three (as minimum) spectral bands from 480 to 560 nm. Operation of the SeaWiFS and MERIS colour scanners ceased in December 2010 and April 2012, respectively, while the MODIS scanner is still working on the Terra and Aqua satellites. In this research, level 2 data (products of standard atmosphere correction at three bands filtered by masks/flags) of SeaWiFS, MODIS (on Terra and Aqua satellites), and MERIS scanners were retrieved for their mission lifetime. The regional algorithm was validated independently for each scanner, based on the adequacy of the algorithm-derived chl-a and a_CDM to in situ-measured data for the same day. The results suggest a satisfactory accuracy of the modified regional algorithm.